YouTube will automatically label photorealistic AI videos, Shorts, too

YouTube moved its AI disclosure labels out of the expanded description and onto the video itself, a shift aimed at making synthetic clips harder to miss as photorealistic AI gets harder to spot. For long-form videos, the label now sits directly below the player. For Shorts, it appears as an overlay on the video, a placement YouTube said would give viewers “the context they need at a glance.”

The company said it would begin automatically labeling videos in May 2026 when its systems detect significant photorealistic AI use, instead of relying only on creators to self-disclose. That step built on a policy YouTube first announced on March 18, 2024, when it required disclosure for realistic altered or synthetic content that could be mistaken for a real person, place, scene or event. Clearly unrealistic animation, fantasy scenes, minor effects and filters, and AI used only for production help such as scripts or captions were left out, and most labels under the earlier system lived in the expanded description unless the video involved sensitive topics like health, news, elections or finance.

The new approach still leaves creators in the loop. YouTube said they can continue disclosing AI use in YouTube Studio, and anyone who thinks a video was misidentified can update the disclosure status. Some labels may stay permanent, including content made with YouTube’s own AI tools such as Veo and Dream Screen, and videos carrying C2PA metadata that indicates they were fully AI-generated. YouTube also said the labels themselves would not affect recommendations or monetization, a safeguard for creators that also limits the policy’s bite.

That is where the enforcement gap becomes plain. YouTube said it uploaded about 20 million videos a day, a scale that makes human review impossible and automated detection essential. The platform had already expanded likeness and deepfake detection in 2026 after earlier testing with celebrities, public figures, politicians and other creators, and in 2024 it partnered with Creative Artists Agency on responsible AI tools for talent. The question now is whether a label below a player, or an overlay on a Short, can do enough in an election year when misleading synthetic video can spread faster than any review queue can catch it.